This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Here's the CVS toad...



The attached tarball contains a program that works for me  :-).
It's purpose is to automatically update and commit generated
files.  This is part one of a two-part process of having the
CVS archive build be able to cope with development environments
that do not have all of the tools.  This is the hard part.

The file "regen.def" defines all the different files and
relationships between generated files and their predecessors.
Those data get munged into a couple of tables that are used
to derive the list of generated files plus all the files
needed to do the generating.  Some generated files depend upon
multiple source files.

Since this is a program that should only ever need to
run on the gcc.gnu.org box, it is not clear where the sources
should live.  I would argue that it implements CVS functionality
and belongs in the source tree for CVS sources.  :-)

REQUIREMENTS FOR RUNNING:

all the generated files must have all the branch tags that
may be used in checking in predecessor files.  This program
*will* regenerate the derived files on the indicated branches.
It determines the branch by grubbing  around in the Entries
file looking for a symbolic branch name sticky tag.

Once this has been done, then just put the program name in
the file:  $CVSROOT/CVSROOT/commitinfo

REGEN.tgz


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]